370 
Journal of Agricultural Research 
Vol. XIV, No. 9 
throughout the entire growing season in 1915 a larger number and greater 
yield of kernels was produced, but the average weight of the kernels was 
not as great as in 1916, when conditions favored a more normal develop¬ 
ment. 
As indicated by the means, the various characters studied responded 
more or less directly to external conditions which prevailed while each 
was making its most rapid development. The number of culms per plant 
and the yield of straw were influenced most by environment during the 
early part of the growing season, and number, yield, and average weight 
of kernels by environment during the latter part of the growing season. 
When growth was 
retarded or stopped by 
environmental condi¬ 
tions, lower yields of 
straw resulted from a 
reduction in the num¬ 
ber, total, or average 
length of culms per 
plant; and lower yields 
of grain, from a re¬ 
duction in the number 
/so 
/*o 
ft 
\ 
k qo 
$ 60 
I 
I a ° 
1 
\ 
/ 
/f 
it 
^7 
1 - . 
J / 
w 
1 
1 
(j 
1 
.1 
1 
1 
1 1 
i 
1 
7 
l 
1 
I 
1 
1 1 
f 
1 
fu 
1 
1 
4r 
1. 
4 
N 
1 
1 
A 
l\ 
\ 
\ 
! 
$ 
■ S 
\L 
\ 
1 \ 
/ X 
1 
3 /o /3 so S3 30 33 4-0 
AH wn or / A/ Sf/LC/G/tAMS 
*3 
Fig. 4.—Graphs showing the frequency distribution of wheat plants 
for average weight of kernels. 1914-1917. 
of kernels; and for 
1914 and 1916 only, a 
lower average weight 
of kernels. The ker¬ 
nels were more or less 
shriveled in 1914 and 
1916. When the grain 
developed normally, as in 1915 and 1917, the lower yield of kernels in 
1917 was accompanied by a higher average weight per kernel. 
standard deviations 
The standard deviations for number of days from planting to second 
leaf are i.928±o. 035 in 1916 and i.330±0.029 in 1917. 
Tor height at second leaf the two highest means are accompanied 
by the highest and lowest standard deviations. Better root development 
in 1917 during the prolonged cool period intervening between the time of 
planting the seed and emergence, which may have permitted the plants 
to begin growth at the surface more nearly at the same time, is a possible ex¬ 
planation of the lower standard deviation accompanying the higher mean. 
The standard deviations for height at six weeks are in the same order 
as the means, with significant differences between any two except those 
in 1914 and 1916. The variability of average height at second leaf and 
at six weeks, measured by the standard deviations, is comparatively low 
during the 4-year period. 
